Slashdot Mirror
PS3 Client for Folding@Home Debuts, ATI GPU Version Soon
eliot1785 writes "Stanford's Folding@Home project is reporting that Sony debuted a Folding@Home client for the PlayStation 3 today in Germany. Researchers hope to use the power of the PS3's Cell processor to greatly expand the number of FLOPS of which their network is capable. F@H also announced today that they will release a client capable of running on ATI graphics processors. With these two new developments, F@H hopes to raise the total power of their distributed computing network to 1-10 petaflops. At the upper end of that target, the network would be faster than any current supercomputer, at least in terms of FLOPS."
Reader TommyBear points out a collection of papers showing scientific advances made by the F@H researchers.
Nice news. I'm sure Sony will make lots of PR capital out of this ala the subject ;)
Will this run on PS3 Linux or natively on the regular OS?
That's like... 10,000,000,000,000,000 instances of taxpayers dollars being wasted! How many more times does this have to flop before it's canceled?
Imagine what would happen if they could also harness Diebold's flops...
Donate free food here
I know they try to justify SETI by saying that you're only donating your "spare cycles". But if you look more carefully, modern computers use 100 Watts to idle and 300 Watts under load. That difference of 200W at US$0.15/kW-hour (my local power rates) comes to THREE HUNDRED DOLLARS PER YEAR of your own money that you donate to the SETI project.
Actually, that's not quite right. It's more like you buy $300 worth of coal every year and burn it with no particular result. Gee, thanks!
that the PS3 is a big FLOPS?
I like the idea of using the GPU for other things, especially since its the most capable parallel processor in most machines and sits almost idle most of the time.
liqbase
x86 continues to get left in the computational dust.
I have a friend who is a very senior engineer at NVidia who has talked about how sick and tired they are of having the boat anchor that is x86 tied to their hardware. And that they would love to just cut out Intel and just run Windows/Linux right on their hardware. Microsoft obviously felt the same way when they dumped Intel and switch to PowerPC with the 360.
The PS3 is supposed to completely support keyboard and mouse, have a full version of Linux sitting on the harddrive, and support homebrew development. If you can download and install normal Linux apps...a graphics programmer dream come true. Even cooler are the plans of Sony coming out with higher end PS3 models with more RAM or Cell chips. A Linux box with a couple gigs of RAM and dual or quad Cells, oh baby.
The summary compares the F@H cluster (or just network) to a supercomputer. That's silly. They shine on very different jobs. "PetaFLOPS" alone reveals little of the real performance. *sigh*
There aren't much details on the ATI version. I'm guessing there's no Nvidia version yet because of the lack of IEEE 854 compliance in viedo cards, so they'd have to create a special version for each video card. But it's pretty neat what you can do with video cards these days besides play video games.
Most people just don't consider the cost. OP never said anyone was forced into it, but then again we all know you're trolling.
Oh, wait, I can't even get a PS3 at all yet, can I, because they haven't even started making them!
What's the fucking point of releasing software for a system that doesn't even exist yet? Cheap PR? Yeah, thanks Sony, it's really great that all these poor people with cancer will soon be cured by your incredible magical non-existent console, but how about you throw in some beef with that bun for a change?
The Cell processor is vastly superior to the "Emotion Engine" and yet, Sony doesn't seem concerned that it'll be used to build supercomputer networks by menacing countries...
;)
sarcasm here, please...
I don't feel like it...
The Broadband Engine in the PS3 has roughly 210 Gflops of power at 3.2Ghz. That is around an order of magnitude more than most people's current Intel desktop PCs. Although that isn't really the full story since it is the memory architecture that makes Cell chips so much more powerful than Intel chips, but that is a whole other, very cool, subject. If even a small percentage of the 100+ million PS3s Sony will sell over the next five years are added to computation pool, the results will be staggering.
if I had points I would mod you +1 insightful.
Carry on.
I know Sony and the PS3 is regularly criticised for their marketing practices - but how much attention is given to what the Cell processor can do?
I haven't seen much discussion about it at all, at the moment I usually see it put forward as slightly more than the 360, but some performance numbers presented way back seemed much more impressive. Can someone write something sensible about what we might expect in terms of pure performance? Will it be the best thing since 7 concurrent simulated nuclear tests run on separate HD-TVs?
Stanford's Folding@Home project is reporting that Sony debuted a Folding@Home client for the PlayStation 3 today in Germany. Researchers hope to use the power of the PS3's Cell processor to greatly expand the number of FLOPS of which their network is capable.
Gee, that's much better (and completely different) than when Saddam was supposedly using playstation 2's to test nuclear weapons. This isn't a planted story by Sony *at all*.
Push Button, Receive Bacon
I'll bet you a million bucks the ps3 has something similar to what tivo uses..
in fact i'll bet you one of the coprocessors on the cell is just like microsoft's custom xbox360 processor, with on-die signing and encryption keys.
Yeah, it'll run linux.. a "signed" distribution of linux which will be the only thing ps3 will run.
why do you think Sony announced it rather than the folding@home coordinators at stanford?
Sony probably only offered the signature because the project submitted everything and assured sony they'd happily stay right under their proprietary lockin thumbs.
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
If one bothers to skim the papers linked, it would appear that the major "product" of the F@H project so far has been to give these particular researchers a lot of knowledge and recognition in the field of folding _in general_.
:) )
Nothing indicates that we are one iota closer to actually applying this knowledge, nor actually performing simulations of real-life _important_ proteins..
Folding @ Home, Seti @ Home.. They seem to exist mainly to advance the careers of their inventors, let overclockers benchmark their machines, and let everyone circlejerk about how they "donate" to "charity"..
(Everyone also conveniently forgets that the pollution caused by the electricity consumed by the elevated cpu usage probably has a worse total impact on the human condition as a whole than any [dubious] future knowledge gains
-Eric
SJW: Someone who has run out of real oppression, and has to fake it.
kill my karma, but you are fucking nuts
if you had a PS3 would you run this in down time?
I'm running folding@home at 2 PCs that runs all night (using electric power at night is more cheap than during the daytime). And it's installed on other 3 PCs, so when I'm only browsing the Internet or so, it uses the unused processor.
So if I buy a PS3, or a Cell personal computer, I'm sure that it is going to run folding@home.
My city: Barcelona.
Somehow I don't think Sony will be too pleased to have the PS3's success measured in flops ;)
--I thought I was wrong once, but I was mistaken.
You're too full of shit for him to reply in full. It'd take all day.
When the posters fear their moderators, there is tyranny; when the moderators fears the posters, there is liberty.
I haven't been following things too closely, but has Nintendo made any moves to make a dev kit available that non-commercial developers can use? If so it would be cool if a version for the Wii could be made for this project.
Jumpstart the tartan drive.
You must be thinking of SETI@Home. That is useless.
Folding@Home does actual science. Thumbs up for those guys!
--cyphertoxin
I would assume whoever wants the massive computation is willing to pay a notable amount to anyone who allows their PS3 to be hooked up to it for a signficant time per month. Perhaps Sony could remind people of the money they "get back" after the high price?
Apology to Ubuntu forum.
But you're still gonna sell the rest of the world arms, right?
when little Jonny's monther made him turn off his game and come upstairs for dinner.
Oh well.
I am very small, utmostly microscopic.
...you're only paying a fixed price per month for your electricity, so it's *just fine* for you to waste as much as you want. Screw the environment, screw the costs you're passing on to others - both financial and ecological. It's not coming out of your wallet, so fuck the World.
Tell me, do you set US foreign policy?
personally, i wonder if anyone's calculated the number of cpu cycles lost due to the alienation of classic seti@home users
i sure do miss the good old days when science wasn't so corrupt and political
ps. pun intended
Now you can use it for both!
"I forgot my mantra."
This is what I get for browsing at -1....
I believe the universal Macintel version is still not out yet. Presumably there's a lot more to it than simply taking the x86 linux sources for the science code and using them in place of the ppc..
How funny would it be that there'd be PS3 work units posted before Macintel?
Still, won't matter to me until the Merom Macbooks come out anyway.....
does this have anything to do with the current topic?
As a long time Folding At Home contributor, I found the following statement to be incorrect:
"Total power of their distributed computing network to 1-10 petaflops. At the upper end of that target, the network would be faster than any current supercomputer, at least in terms of FLOPS."
The fastest supercomputer in the world is IBM's Blue Gene/L, which clocks in at 280.6 teraflops http://www.top500.org/. The distributed network of Folding at Home is currently 171.2 teraflops http://folding.stanford.edu/stats.html/ as measured by sustained contributions from active members. The Folding At Home network is already the second most powerful, if ranked by the Top 500 list criteria. At any range from 1-10 petaflops, the Folding at Home network would be more powerful than the most powerful supercomputer in the world.
How many Flops is a Beowulf Cluster of Alaskan Diebold machines...?
Find Japanese addresses in English on Google Maps Japan: http://diddlefinger.com/
You guys rock!
...and they call *me* a troll.
Instead of opposing this viewpoint with your own, or attempting to explain or debate aspects of it or arguments against it, you insult the poster.
*laughing*
But can you fly around the molecules by waving the controller around like you're flying a futuristic aircraft?
Reviewing just the first hour of video games.
That's the sound of all those NDA's being lifted.
;)
Only about 150 still in effect now.
Wait till you see what's next
.
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
I think it would be cool if Windows' screensaver were programmed to automatically join this thing, no matter which screensaver the user chose. That way, there would be zillions and zillions of petaflops available to this process, without having to do anything special to achieve it.
ahaaa! His second post was modded "flamebait" too! It's no wonder Slashdot is on the wane. The mods here have staked out their niche and cling to it just to push their little agendas. Have fun while it lasts...
How nearsighted of them ;) No support for PPC64 at all? I even tried building Wine on an 8-way POWER5 machine to run the Windows 32-bit binary under, and that didn't work either.
So how about it? When will we see a PPC/PPC64 Linux binary of Folding@Home? Where is the source, Luke? I'll build it myself!
Spoken like a true ignorant American who doesn't know shit about the rest of the world, or, indeed, about the politics of the United States itself.
The entire planet opposed us for a reason, or rather, many reasons. Even the countries that "supported" us, did so against the wishes of the majority of their respective populations, and only to win our favor.
Just look at the disaster we've created in Iraq. All we've done is destroy infrastructure, further damaging the quality of life of Iraqis, and even worse, removed the keystone preventing civil war -- yes, Saddam Hussein and his government, our former ally. Oh... and built an oil pipeline...
Sticking feathers up your butt does not make you a chicken - Tyler Durden
For no reason other than because I'm evil, I present to everyone the following back-of-the-napkin/sources-from-wikipedia analysis:
There was an article a while back about game console power consumption, but rather than dig that up, I'll assume a PS3 will average 200 Watts while cranking away on proteins. It's a good, round number. And I'll assume that I'd spend an hour per day actually playing games. Electricity in my area costs about $0.08/kW-hr.
0.2 kW * 23 hr/day * 365 day/year = 1679 kW-hours/year
1679 kW-hr/year * $0.08/kW-hr = $134.32/year for electricity to fold imaginary proteins. Ouch.
And for those worried about C02, 1679 kW-hr is 6,044,400 kJ, which is the energy equivalent to 46 gallons of gasoline (efficiency of conversion not accounted for). Alternately, assuming your electricity comes from a natural gas (CH4 ~ 891 kJ/mol) plant operating at 40% efficiency, one year of folding on your PS3 would release 746 pounds of CO2 (plus 1220 pounds of water vapor).
Gee, aren't numbers fun? In the fight to cure cancer, you actually end up breaking the bank and destroying the planet. That sucks.
I probably really shouldn't have posted that. I'm going to give all the idealistic, penny-pinching, obsessive-compulsive, environmentalist slashdot readers a complex.
I, for one, welcome our petafloping folding overlords.
Do anybody know the actual power consumption when CELL is running at full speed and normal working speed (if it exist and is not full speed)? While grid computing is definitely an attractive idea, running this on PS3 means it runs on the background when you are gaming/watching, and on the foreground when you rest? Considering the electricity consumption of XBox 360, PS3 is not likely to be much more power-saving. People around me won't like to see such a heat-emitting and perhaps whining monster turned on from dawn till dawn.
Differential pricing for peak and off peak electricity is fairly common here in the UK. It is called Economy 7 because we get 7 hours of cheaper electricity overnight. I pay 8.77p/KWh in the day but only 3.6p/KWh at night. Most installations of this type use the off peak electricity to heat up night storage heaters and hot water tanks.
Night storage heaters are particularly crap because you have to know in advance what the weather will do. If you turn it off overnight and the weather is cold the next day then you have to use expensive electricity to stay warm. If you leave it on and the weather is warm, you roast and there is nothing you can do about it. Storage heaters often make the bedroom too hot overnight but that is necessary in order to have heat the following evening.
A latent existence
I think submitter is a folding@home user like me and wanted to share these great news with Slashdot "technical community".
4 9
You see the result as half of the posters say PS3 has a bad performance and other half says folding@home, a scientific project is sort of "fake" thing.
Please don't "share" any folding@home news with this user profile from now on.
As folding@home is a scientific computation , if people get wrong impression from these "know everything" geeks they will simply delete folding@home. That _is_ a problem. Folding@home sends and reserves a data to a single machine and expect it e.g. 80 days later. If the user gets that data, nobody touches that data until it expires. So there can't be a worse thing to do against that project unless you start DOS attack to Stanford servers.
Lets make it more clear. Here is one of proteins in my quad G5 machine gets computed while I write this message.
" issue: Wed Aug 16 14:54:13 2006; begin: Wed Aug 16 14:54:17 2006
expect: Sat Aug 26 00:56:54 2006; due: Thu Nov 30 13:54:17 2006 (106 days)"
"p2128_ww2128"
So, if I am end user who gets folding@home and doesn't know how project works, I see couple of geeks saying it is a bad, useless thing, I uninstall folding@home yes? That portion of data BELONGS to my computer and there is no way Stanford will figure until 106 days. Another guy could be folding it.
An article discussing seti@home and if it is needed on WSJ has made to slashdot
http://slashdot.org/article.pl?sid=06/06/28/05562
Of course, there were AC posting like "I am in medical industry, nobody needs a thing like this" and Doctor Vijay Pande had to "reply" using Stanford forums giving up his real work to reply some fake medical worker. I believe because of that +5 interesting AC post (yes, a while!) many people uninstalled folding@home really wasting work units AND the CPU time they spent so far.
Other real scientists had to spare their time to massively stupid predictions like "IP will be locked for 70 years" too.
It is NOT a game, it is not sort of FAKE thing, people of Stanford and GROMACS like organisations spare thousands of hours of professional work and users donate their IDLE cpu time to project which has impact on REAL science and REAL medical work.
I really don't care if you hate PS3 processor or memory architecture, just stay hell out of hurting a scientific project running for years. We (folding community) really don't need your types watts or processor power at all.
BTW same applies to BOINC etc too.
1) The whole thread is not about having an OS running ON a bare naked GPU, but the fact that modern GPU are massive and parallel floating-point units that could be harnessed to do wonderful specialised calculation (things that can be parallelized without too much precision needed, some thing that some BOINC projects like Folding @ home to nicely), and on the other hand nVidia complaining that while they have all this interesting opportunities, they are forced, due to market share, to spend most of their time supporting a brainfucking stupid OS (most of the recent Windows version that still needs GPU drivers) legacy ridden architecture (x86. Even recent chips like the Pentium4 and the Core are still backward compatible with the tiny-winy number of registers the banked memory model and some BCD (maths in base 10) of the original 8088/8087. At least, more of the legacy is now handled in microcode (register renaming) and newer architecture like AMD64 got some of those stupidity removed, specially when running in 64 mode. That's also one of the main reason why the Itanic failed : every one is tied to this old crap architecture, and though crappy it is nobody could suddenly switch to some other more modern architecture. Even if nVidia would like to spend more time with more custom tailored achitecture ).
2) We don't speak about architecture related restriction, but about market share. Most of the company have to spend most of the work time supporting the main dog (Windows on Intel). Yes, we know that you could plus those card into anything with a PCI, an AGP, or an Express connecter. But how many people are pluging those cards into a sparc ? Not many. Windows is still the main market that must be supported (because of the huge amount of gamers that stay there) and the linux users and/or exotic hardware user have just no luck. (And that, coupled with the lack of collaboration from both ATI and nVidia to develop good open-source drivers makes me angry. nVidia users are out of luck for OSS 3D, ATI must rely on reverse-enginered stuff ).
3) Cell for your information, the Cell processor includes a bunch of highly specialised co-processors (dedicated hardware good for vector and floating-point) NEXT to a general purpose CPU Unit (PowerPC based, as far as I know). The CPU parts runs the OS, the cells only serves the dedicated math tasks. Think of it exactly like what is done with GPUs-turned-into-general-purpose-math-units, or the different (programmable) physics engine, etc... the cell are just here to offer a programmable vector/flop facility to the software running on the PowerPC part.
On the other hand given the crap that was able to run OS (xx-DOSes on 8088) some primitive OS *may* be ran on the cell part...
4) Microsoft switched from x86 to PPC, because it's their frickin' hardware, where they control everything and are free to choose whatever solution they want, because the consumer buys a 'XBox', he doesn't care if there's a "Intel Inside" or "powered by Windows" (...of in case of DreamCast : "compatible with Windows" for the few applications that did use it). The PPC switch gave them opportunities they found interesting (less knowledge around hackers, so they hoped slower breakin in. custom built hardware meant getting free of legacy x86 architecture that still somewhat plagued the XBox 1), etc... Although the x86 for the XBox 1 was a logical move : for a company that mostly works with Windows PC, making a gaming console that is basically a customized Windows running on a simplified PC both has an easier learning curve and is cheaper to produce (due to the high availability of PC hardware around). Basically XBox 1 was an experiment in the console field using a PC with "XBox" written on it, were as XBox 360 is more trying to develop in complete freedom a gennuine console architecture.
Apple on the other hand switched specifically to Intel for mostly a marketing reason : the computers today are a commodity, and Apple is trying to attention span from,
- people
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
I seem to recall reports of PS1's and PS2's dying when Final Fantasy was run nonstop for days. So how long will a PS3 be able to crank away on protein folding before it dies from overwork? Just make sure you've got adequate cooling, maybe?